Krypf’s Diary

暇なときに書く

Errors in Geometry, Topology and Physics

Table of Contents

Errors in Nakahara's book

Geometry, Topology and Physics by Mikio Nakahara can be downloaded from the following online page.

(GradMikio Nakahara-Geometry, topology, and physics-Institute of Physics Publishing.pdf
http://staff.ustc.edu.cn/~hyx/0911/(GradMikio%20Nakahara-Geometry,%20topology,%20and%20physics-Institute%20of%20Physics%20Publishing.pdf

If you want to buy the published version, there is an Amazon page and you can get it anytime.

Amazon | Geometry, Topology and Physics, Second Edition (Graduate Student Series in Physics) | Nakahara, Mikio | Differential Geometry
https://www.amazon.co.jp/-/en/Mikio-Nakahara/dp/0750306068

1. Example 9.2.

Right after the equation (9.19), there is the following sentence:

Let  W=W^{\nu }\partial /\partial y^{\nu } be a vector of  T _p N and  V=V^{\mu}\partial /\partial x^{\mu} be the corresponding vector of  T _{f ( p)} M .

This is wrong, although physicists and physics students don't care about such a thing.

Two vector fields  W and  V must be actually just vectors, that is,

 W=W^{\nu}\left( \dfrac{\partial }{\partial y^{\nu }}\right) _{p}

and

 V=V^{\mu }\left( \dfrac{\partial }{\partial x^{\mu}}\right) _{f\left( p\right) } .

Obviously, the author writes "a vector of  T _p N " so  W is not a vector field. Similarly,   V is a vector.

2. Section 9.2

Just below the equation (9.10), the author writes the definition of a section.

He writes the set of sections of a fiber bundle  (E,\pi, M, F) as  \Gamma(M,F) ,

which is incorrect. A section maps a point on the base space to an element of the total space.

In fact, the following example is the sections on  M ,  \Gamma(M, TM ) , where the tangent bundle  TM is not a fiber but the total space.

Correctly, the formula is

 \Gamma(M,E) .

cf. Section (fiber bundle) - Wikipedia
https://en.wikipedia.org/wiki/Section_(fiber_bundle)

3. Section 10.1.3

Below Figure 10.2 in section 10.1.3, he writes

where we have noted that  \sigma _{i\ast }X\in T_{\sigma _{i}}P

but this is wrong. That's because  \sigma _{i} is not a point on  P .

In this section,  P is a principal bundle. In other words, given a fiber bundle  (E, \pi, M, F, G) , the total space is  E \equiv P and being a principal bundle gives  F = G by definition.

 \circ Projection

In section 9.2.1, a differentiable fiber bundle is defined and the projection is important in this case:

(iv) A surjection  \pi :E\rightarrow M called the projection. The inverse image  \pi ^{-1}\left( p\right) =F_{p}\cong F is called the fiber at  p .

A fiber  F_{p} at  p is a subset of  E and just needed to be diffeomorphic to the typical fiber   F . If we rewrite it into a mathematical formula,

 F \cong F_{p} = \pi^{-1} [\lbrace p  \rbrace] \subseteq E .

 \circ Section

Def.

Let  (E,\pi, M,F,G) be a fiber bundle. A section (or a cross section)  s is defined as

  1.  s:M\rightarrow E
  2.  \pi \circ s=\mathrm{id}_{M}
  3. smooth (differentiable fiber bundle) or homeomorphic (topological fiber bundle).

By definition, the value of a section at  p ,  s(p) , is an element of a fiber over  p ,  F_{p} = \pi^{-1}[\lbrace p \rbrace] , i.e.

 s(p) \in F_{p} \cong F and  F_{p}  \subseteq E .

The set of (global) sections on  M is denoted by  \Gamma(M,E) (not  \Gamma(M,F) ).

Here, a local section  \sigma _{i} is an element of  \Gamma \left( U_{i},E\right) = \Gamma \left( U_{i},P\right) , where  U_{i} is an open set of the base space  M .

Let  p be a point on  M , then  \sigma _{i}(p) \in F_{p} \subseteq E \equiv P . Therefore, the correct sentence is

 \sigma _{i\ast }X\in T_{\sigma _{i}(p)}P .

Summary

I found three errors in the contents of Nakahara's book.

  • Not vector fields, but vectors.
  • Sections are the total space  E -valued functions.
  • Misprint about a tangent space.